What changed
AWS Quick, an AI-powered service for automating tasks, now supports the use of agents and skills to perform repetitive tasks such as building status updates, writing release notes, and managing projects. The service enables users to interact with it in natural language, and agents process requests against connected data sources and applications.
What this means in practice
Users can work with agents in chat or let them run on their own using routines. Agents are configurable with their own instructions, knowledge sources, and connectors, allowing them to answer questions and perform tasks for the user. Quick is the default built-in agent available to all users, which adapts to the resources the user has access to, such as spaces, knowledge bases, dashboards, connectors, and skills.
Custom agents can be created for specific use cases, such as an HR onboarding assistant that answers new-hire questions from an HR team’s space. These agents can be tested, published, and shared within an AWS Quick account.
Skills are reusable sets of instructions that teach agents how to perform specific tasks, and they are built using the Agent Skills specification, an open standard developed by Anthropic.
Limitations and availability
Skills are folders containing a Markdown file with metadata and optional scripts. The description and name of a skill are crucial for triggering it, and Quick uses progressive disclosure to load skills into the agent’s prompt. The AWS Quick desktop version, which was in preview in April 2026 and became generally available in September 2026, uses this specification. The service supports various connectors for external applications and tools, with cloud connectors available on all devices and local connectors requiring a machine to be on and connected.
